[PATCH] D53543: [analyzer] MallocChecker: pr39348: Realize that sized delete isn't a custom delete.

[Umann Kristóf via Phabricator via cfe-commits]

Szelethus added inline comments.


================
Comment at: lib/StaticAnalyzer/Checkers/MallocChecker.cpp:727
 
-  // Skip all operator new/delete methods.
-  if (isa<CXXMethodDecl>(FD))
-    return false;
-
-  // Return true if tested operator is a standard placement nothrow operator.
-  if (FD->getNumParams() == 2) {
-    QualType T = FD->getParamDecl(1)->getType();
-    if (const IdentifierInfo *II = T.getBaseTypeIdentifier())
-      return II->getName().equals("nothrow_t");
-  }
-
-  // Skip placement operators.
-  if (FD->getNumParams() != 1 || FD->isVariadic())
-    return false;
-
-  // One of the standard new/new[]/delete/delete[] non-placement operators.
-  return true;
+  // This is standard iff it's not defined in a user file.
+  SourceLocation L = FD->getLocation();
----------------
typo: if


================
Comment at: test/Analysis/NewDelete-custom.cpp:3-4
 // RUN: %clang_analyze_cc1 -analyzer-checker=core,cplusplus.NewDelete,cplusplus.NewDeleteLeaks,unix.Malloc -std=c++11 -analyzer-config c++-allocator-inlining=false -DLEAKS=1 -fblocks -verify %s
 // RUN: %clang_analyze_cc1 -analyzer-checker=core,cplusplus.NewDelete,unix.Malloc -std=c++11 -DALLOCATOR_INLINING=1 -fblocks -verify %s
 // RUN: %clang_analyze_cc1 -analyzer-checker=core,cplusplus.NewDelete,cplusplus.NewDeleteLeaks,unix.Malloc -std=c++11 -DLEAKS=1 -DALLOCATOR_INLINING=1 -fblocks -verify %s
 #include "Inputs/system-header-simulator-cxx.h"
----------------
Do we still need `-DLEAKS=1` and ` -DALLOCATOR_INLINING=1`?


https://reviews.llvm.org/D53543